When Dinesh Said AI Can’t Solve His Problems
Last week was about Dinesh and his line that stopped me in my tracks:
“AI is not capable of solving my problems.”
This week I saw what he meant.
AI didn’t fail. It just showed its limits.
We spent most of the week inside Cursor and ChatGPT, connecting systems, refining logic, and setting up automation flows that used to take days. It’s fast, clean, and impressive.
But it also felt like we were living with a quiet co-worker who never really understands the bigger picture.
I have started to see AI as a mirror more than a machine. It reflects what we already know, what we’ve done before, what we’ve written and built.
– It doesn’t imagine.
– It doesn’t dream.
– It predicts.
(And makes me reallllllly lazy haha)
And that’s fine, as long as you know what you want.
Clients are beginning to expect AI to handle everything.
Scope, plan, test, deploy.
The conversation has shifted from “how fast can you build this?” to “can’t AI just do it?”
And that’s where it gets tricky. Because yes, AI can do a lot. But not enough.
The projects that actually work are the ones where people step in, where someone questions an output, where intuition decides what the algorithm can’t.
We’re learning that the future of building products is not human versus AI.
It’s human through AI.
Our job is to shape it, not follow it.
To use AI as an amplifier, not as a replacement.
To be the ones who still ask “why?” when AI just says “here’s how.”
Week 43 taught me something simple.
AI isn’t here to take our jobs. It’s here to test our value.
The real question isn’t what it can do.
It’s what we still choose to do better.
Last week was about Dinesh and his line that stopped me in my tracks:
“AI is not capable of solving my problems.”
This week I saw what he meant.
AI didn’t fail. It just showed its limits.
We spent most of the week inside Cursor and ChatGPT, connecting systems, refining logic, and setting up automation flows that used to take days. It’s fast, clean, and impressive.
But it also felt like we were living with a quiet co-worker who never really understands the bigger picture.
I have started to see AI as a mirror more than a machine. It reflects what we already know, what we’ve done before, what we’ve written and built.
– It doesn’t imagine.
– It doesn’t dream.
– It predicts.
(And makes me reallllllly lazy haha)
And that’s fine, as long as you know what you want.
Clients are beginning to expect AI to handle everything.
Scope, plan, test, deploy.
The conversation has shifted from “how fast can you build this?” to “can’t AI just do it?”
And that’s where it gets tricky. Because yes, AI can do a lot. But not enough.
The projects that actually work are the ones where people step in, where someone questions an output, where intuition decides what the algorithm can’t.
We’re learning that the future of building products is not human versus AI.
It’s human through AI.
Our job is to shape it, not follow it.
To use AI as an amplifier, not as a replacement.
To be the ones who still ask “why?” when AI just says “here’s how.”
Week 43 taught me something simple.
AI isn’t here to take our jobs. It’s here to test our value.
The real question isn’t what it can do.
It’s what we still choose to do better.